摘要
风力发电机组是风电系统的核心组件。发电机的效率和稳定性直接影响风力发电机组的整体发电能力和可靠性,对风电场的经济效益至关重要。提出了一种基于SCADA数据和XGBoost算法的风电机组发电机故障诊断方法。通过采集风力发电机组的实时SCADA系统数据,包括风速、功率输出、温度等关键指标,并利用这些数据作为模型输入特征。采用极致梯度提升树(XGBoost)算法进行故障诊断模型的训练和测试。XGBoost作为一种高效的梯度提升算法,具有处理高维特征和大规模数据的能力,并且在分类问题中表现出色。在模型开发过程中,对SCADA数据进行了预处理,包括数据清洗和特征选择。通过交叉验证方法对XGBoost模型的参数进行了优化,以提高模型的泛化能力和准确性。实验结果表明,基于SCADA数据的XGBoost的诊断模型能够有效识别出的发电机故障,其准确率达到了93.14%,精确率、召回率和F1分数指标均高于随机森林和K-邻近模型。
The wind turbine generator is the core component of the wind power system.Its efficiency and stability directly affect the overall power generation capacity and reliability of wind turbines,and are crucial to the economic benefits of wind farms.This proposes a fault diagnosis method for wind turbine generators based on SCADA data and XGBoost algorithm.By collecting real-time SCADA system data of wind turbines,including key indicators such as wind speed,power output,and temperature,and using these data as input features for the model.The XGBoost algorithm was used to train and test the fault diagnosis model.XGBoost as an efficient gradient boosting algorithm,had the ability to handle highdimensional features and large-scale data,and performs well in classification problems.During the model development process,was the SCADA data preprocessed,including data cleaning and feature selection.The parameters of the XGBoost model were optimized through cross validation to improve its generalization ability and accuracy.The experimental results show that the XGBoost diagnostic model based on SCADA data can effectively identify generator faults,with an accuracy of 93.14%.The accuracy,recall,and F1 score indicators are all higher than those of random forest and K-nearest neighbor models.
作者
施德华
韩增涛
孟井煜枫
吴博阳
汪浩然
蔡逸波
SHI Dehua;HAN Zengtao;MENG Jingyufeng;WU Boyang;WANG Haoran;CAI Yibo(Windey Energy Technology Group Co.,Ltd.,Hangzhou 310000,China)
出处
《微特电机》
2025年第5期61-64,70,共5页
Small & Special Electrical Machines
